README.md

metadata

language:
  - ko
license: cc-by-nc-4.0
model-index:
  - name: LDCC-SOLAR-10.7B
    results:
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: AI2 Reasoning Challenge (25-Shot)
          type: ai2_arc
          config: ARC-Challenge
          split: test
          args:
            num_few_shot: 25
        metrics:
          - type: acc_norm
            value: 67.32
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=LDCC/LDCC-SOLAR-10.7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: HellaSwag (10-Shot)
          type: hellaswag
          split: validation
          args:
            num_few_shot: 10
        metrics:
          - type: acc_norm
            value: 88.11
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=LDCC/LDCC-SOLAR-10.7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MMLU (5-Shot)
          type: cais/mmlu
          config: all
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 66.83
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=LDCC/LDCC-SOLAR-10.7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: TruthfulQA (0-shot)
          type: truthful_qa
          config: multiple_choice
          split: validation
          args:
            num_few_shot: 0
        metrics:
          - type: mc2
            value: 68.85
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=LDCC/LDCC-SOLAR-10.7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: Winogrande (5-shot)
          type: winogrande
          config: winogrande_xl
          split: validation
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 83.66
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=LDCC/LDCC-SOLAR-10.7B
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: GSM8k (5-shot)
          type: gsm8k
          config: main
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 53.6
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=LDCC/LDCC-SOLAR-10.7B
          name: Open LLM Leaderboard

Model Card for LDCC-SOLAR-10.7B

Developed by : Wonchul Kim (Lotte Data Communication AI Technical Team)

Hardware and Software

• Hardware: We utilized an A100x4 * 1 for training our model
• Training Factors: We fine-tuned this model using a combination of the DeepSpeed library and the HuggingFace TRL Trainer / HuggingFace Accelerate

Method

• This model was trained using the learning method introduced in the SOLAR paper.

Base Model

• yanolja/KoSOLAR-10.7B-v0.1 (This model is no longer supported due to a tokenizer issue.)

Caution

• If you want to fine-tune this model, it is recommended to use the tokenizer.json and tokenizer_config.json files from revision v1.1.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	71.40
AI2 Reasoning Challenge (25-Shot)	67.32
HellaSwag (10-Shot)	88.11
MMLU (5-Shot)	66.83
TruthfulQA (0-shot)	68.85
Winogrande (5-shot)	83.66
GSM8k (5-shot)	53.60